ietf-822
[Top] [All Lists]

more on Real and un-Real headers

1991-09-20 16:50:12
I believe that the recent discussion amounts to this: As long as we
specify specific headers or phrases in specific headers which have no
non-display semantics then there is no theoretical problem with
changing the display semantics of only those components [not Received
lines, not X-Fruit-of-the-Day]. The problem is that it would increase
the amount of quoting and quoting is not well handled.

Actually the un-Real proposal has no impact on intermediate mail agents.
We could even do it without standardization. We define

   X-Content-header-charset: mnemonic/7bit

Co-operating users [doesn't even have to be co-operating hosts] can generate
messages which conform to this, and use mail readers which change the way
they display certain headers when they get a message with that header.
People who don't understand the header will still see special characters
in a semi-readable format.

Now removing the need for private agreements is one of the aims of rfc-xxxx
so it would be much better to standardize.

It would also be nice to change the quoted-printable hex lead-in character
(currently ":") to something that is not in the rfc-822 special character
list. Perhaps "?" or "~". Now we could have

   Content-header-charset: anything/quoted-printable

and use quoted-printable quoting to remove the need for rfc-822 quoting,
if we agree that the latter is not well handled.

Note that the proposal in this form is not restricted to mnemonic though
that is the obvious choice. The confusion here was due to my original
suggestion that we make Content-header-charset: mnemonic/7bit the default
for messages with a Content-type header. That suggestion had no support.

Bob Smart

<Prev in Thread] Current Thread [Next in Thread>
  • more on Real and un-Real headers, Bob Smart <=